Major Lapse: Trump’s Acting Cyber Chief Allegedly Uploaded Sensitive Government Files To ChatGPT

The acting head of the U.S. Cybersecurity and Infrastructure Security Agency (CISA) is under scrutiny for reportedly uploading sensitive “For Official Use Only” government documents to public AI platforms, including ChatGPT. The alleged actions bypassed standard security protocols and automated alerts, creating a potential risk of unauthorized data exposure.

Security Alerts Triggered
Reports indicate that the official’s actions prompted multiple warnings from the government’s automated security systems, which are designed to prevent accidental or malicious leaks of sensitive files. While the acting chief was reportedly allowed limited use of ChatGPT, other staff members were restricted from sharing documents on the platform.

Risk of Data Exposure
Uploading internal documents to large language models (LLMs) carries significant cybersecurity risks. AI platforms can retain and learn from user-submitted data, raising the possibility that confidential information could appear in responses to other users. Even files considered “unclassified” may contain operational details that must remain confidential under government policy.

Previous Controversies and Internal Action
The official’s appointment dates back to the Trump administration. Earlier reports indicated a failed counterintelligence polygraph test, later invalidated by the Department of Homeland Security. Following the incident, six staff members under the official’s supervision were suspended from accessing classified information.

CISA Statement
A CISA spokesperson stated that the official’s ChatGPT usage was limited in scope and duration. The Department of Homeland Security is conducting an investigation to determine the extent of any potential damage caused by the uploads.

Expert Warnings on AI and Sensitive Data
Cybersecurity experts stress that both government officials and private users should exercise extreme caution when uploading sensitive documents—including financial records, medical files, identification documents, or legal agreements—to AI platforms. AI providers often review chat data to improve system performance, which could inadvertently expose confidential information. Once incorporated into AI models, sensitive data may also resurface in responses to unrelated queries.

The incident highlights the ongoing challenges of integrating AI tools in secure environments, emphasizing the need for strict protocols and awareness of potential risks.